bedrock-agentcore-starter-toolkit 0.1.13__tar.gz → 0.1.14__tar.gz

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/CHANGELOG.md

@@ -1,5 +1,16 @@
 # Changelog
 
+## [0.1.14] - 2025-09-25
+
+### Changes
+
+- Fix: Runtime configure function now sets CodeBuild execution role from --code_build_execution_role parameter (#184) (7d7dffd)
+- docs: Generate llm.txt via mkdocs-llmstxt (#213) (6459979)
+- fix: llm.txt typo (#210) (f48ae5e)
+- docs: Add llm.txt and file on runtime deployment (#202) (90dac4b)
+- fix: correct pyproject.toml installation in subdirectories (#207) (ea01c65)
+- chore: bump version to 0.1.13 (#205) (86269a9)
+
 ## [0.1.13] - 2025-09-24
 
 ### Changes

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: bedrock-agentcore-starter-toolkit
-Version: 0.1.13
+Version: 0.1.14
 Summary: A starter toolkit for using Bedrock AgentCore
 Project-URL: Homepage, https://github.com/aws/bedrock-agentcore-starter-toolkit
 Project-URL: Bug Tracker, https://github.com/aws/bedrock-agentcore-starter-toolkit/issues

bedrock_agentcore_starter_toolkit-0.1.14/documentation/docs/mcp/agentcore_runtime_deployment.md

@@ -0,0 +1,48 @@
+### Build your first agent or transform existing code
+
+#### Prerequisites & Environment Setup
+- **Environment**: Set up Python 3.10+ and virtual environment - [Environment Setup](https://aws.github.io/bedrock-agentcore-starter-toolkit/user-guide/runtime/quickstart.html#step-0-setup-folder-and-virtual-environment)
+
+#### Step 1: Install Dependencies & Create Agent Code
+- **New Agents**: Install AgentCore packages and create your agent using hello world strands agents  [Installation & Creation Guide](https://aws.github.io/bedrock-agentcore-starter-toolkit/user-guide/runtime/quickstart.html#step-1-install-and-create-your-agent)
+- **Existing Agents**: Transform your current agent code to work with AgentCore - [Framework Integration Examples](https://aws.github.io/bedrock-agentcore-starter-toolkit/examples/index.html)
+
+#### Step 1.1: For strands refer to following documentation 
+- [Welcome](https://strandsagents.com/latest/documentation/docs/index.md)
+- [Amazon Bedrock](https://strandsagents.com/latest/documentation/docs/user-guide/concepts/model-providers/amazon-bedrock/index.md)
+- [Amazon Bedrock AgentCore](https://strandsagents.com/latest/documentation/docs/user-guide/deploy/deploy_to_bedrock_agentcore/index.md)
+1. Import strands agent - `from strands import Agent`
+2. Create an agent with default settings - agent = Agent()
+3. Ask the agent a question - agent("Tell me about agentic AI")
+
+#### Step 1.2 - Transforming agent code for agentcore 
+- **Agent code**: Always use these [code patterns](https://aws.github.io/bedrock-agentcore-starter-toolkit/user-guide/runtime/overview.html#agent-development-patterns) for agent code.
+- **AgentCore Wrapper**: Use  [bedrock-agentcore](https://aws.github.io/bedrock-agentcore-starter-toolkit/user-guide/runtime/overview.html#what-is-the-agentcore-runtime-sdk) wrappers to implement runtime service contract.
+1. Import the Runtime App with from bedrock_agentcore.runtime import BedrockAgentCoreApp
+2. Initialize the App in your code with app = BedrockAgentCoreApp()
+3. Decorate the invocation function with the @app.entrypoint decorator
+4. Create a requirements.txt file with needed packages. Note: if strands-tools is detected, the correct library to add is strands-agents-tools
+5. Let AgentCore Runtime control the running of the agent with app.run() 
+
+#### Step 2: Local Development & Testing (Optional)
+- **Local Testing**: Run and test your agent locally before deployment - [Local Testing Guide](https://aws.github.io/bedrock-agentcore-starter-toolkit/user-guide/runtime/quickstart.html#step-2-test-locally)
+1. Start the agent using python <my_agent.py>
+2. # Test it (in another terminal)
+    curl -X POST http://localhost:8080/invocations \
+    -H "Content-Type: application/json" \
+    -d '{"prompt": "Hello!"}'
+3. Stop the agent.
+
+#### Step 3: Deploy to AgentCore using CLI
+Refer to https://aws.github.io/bedrock-agentcore-starter-toolkit/api-reference/cli.html
+1. Install cli with 'pip install bedrock-agentcore-starter-toolkit'
+2. **Configuration**: Use AgentCore CLI to configure your agent for deployment.
+    ```agentcore configure --entrypoint converted_agentcore_file.py --non-interactive```
+3. **Deployment**: Launch your agent to AWS with automatic resource creation.
+    ```agentcore launch```
+4. **Invocation**: agentcore invoke '{"prompt": "Hello"}' Test your deployed agent using the CLI or API calls
+
+#### Step 4: Troubleshooting & Enhancement
+- **Common Issues**: Resolve deployment and runtime issues - [Troubleshooting Guide](https://aws.github.io/bedrock-agentcore-starter-toolkit/user-guide/runtime/quickstart.html#troubleshooting)
+- **Advanced Features**: Add memory, authentication, and gateway integrations - [Next Steps](https://aws.github.io/bedrock-agentcore-starter-toolkit/user-guide/runtime/quickstart.html#next-steps)
+- **Monitoring**: Set up observability and monitoring for production agents

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/documentation/docs/user-guide/runtime/overview.md

@@ -19,7 +19,7 @@ if __name__ == "__main__":
 
 ```bash
 # Configure and deploy your agent
-agentcore configure --entrypoint my_agent.py
+agentcore configure --entrypoint my_agent.py --non-interactive
 agentcore launch
 agentcore invoke '{"name": "Alice"}'
 ```

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/documentation/mkdocs.yaml

@@ -127,6 +127,28 @@ plugins:
             docstring_style: google
             show_root_heading: true
             show_source: true
+  - llmstxt:
+      markdown_description: "Amazon Bedrock AgentCore enables you to deploy and operate highly capable AI agents securely, at scale. It offers infrastructure purpose-built for dynamic agent workloads, powerful tools to enhance agents, and essential controls for real-world deployment. AgentCore services can be used together or independently and work with any framework including CrewAI, LangGraph, LlamaIndex, and Strands Agents, as well as any foundation model in or outside of Amazon Bedrock, giving you ultimate flexibility."
+      sections:
+        "User Guide":
+          - "mcp/agentcore_runtime_deployment.md"  # replace with user-guide/runtime/quickstart.md
+          - "user-guide/runtime/overview.md": "Runtime service overview"
+          - "user-guide/runtime/permissions.md": "Runtime permissions guide"
+          - "user-guide/runtime/async.md": "Runtime async processing"
+          - "user-guide/runtime/notebook.md": "Runtime notebook integration"
+          - "user-guide/gateway/quickstart.md": "Gateway quickstart guide"
+          - "user-guide/memory/quickstart.md": "Memory quickstart guide"
+          - "user-guide/identity/quickstart.md": "Identity quickstart guide"
+          - "user-guide/builtin-tools/quickstart-browser.md": "Browser tool quickstart"
+          - "user-guide/builtin-tools/quickstart-code-interpreter.md": "Code interpreter quickstart"
+          - "user-guide/observability/quickstart.md": "Observability quickstart guide"
+          - "user-guide/import-agent/quickstart.md": "Import agent quickstart"
+          - "user-guide/import-agent/overview.md": "Import agent overview"
+          - "user-guide/import-agent/configuration.md": "Import agent configuration"
+        "Examples":
+          - examples/*.md
+        "API Reference":
+          - api-reference/*.md
 
 extra:
   social:

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "bedrock-agentcore-starter-toolkit"
-version = "0.1.13"
+version = "0.1.14"
 description = "A starter toolkit for using Bedrock AgentCore"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -165,5 +165,6 @@ dev = [
     "mkdocs~=1.6.1",
     "mkdocs-macros-plugin~=1.3.7",
     "mkdocs-material~=9.6.12",
+    "mkdocs-llmstxt~=0.1.0",
     "mkdocstrings-python>=1.16.10,<1.19.0",
 ]

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/src/bedrock_agentcore_starter_toolkit/cli/runtime/commands.py

@@ -167,6 +167,7 @@ def configure(
     entrypoint: Optional[str] = typer.Option(None, "--entrypoint", "-e", help="Python file with BedrockAgentCoreApp"),
     agent_name: Optional[str] = typer.Option(None, "--name", "-n"),
     execution_role: Optional[str] = typer.Option(None, "--execution-role", "-er"),
+    code_build_execution_role: Optional[str] = typer.Option(None, "--code-build-execution-role", "-cber"),
     ecr_repository: Optional[str] = typer.Option(None, "--ecr", "-ecr"),
     container_runtime: Optional[str] = typer.Option(None, "--container-runtime", "-ctr"),
     requirements_file: Optional[str] = typer.Option(
@@ -268,6 +269,7 @@ def configure(
             agent_name=agent_name,
             entrypoint_path=Path(entrypoint),
             execution_role=execution_role,
+            code_build_execution_role=code_build_execution_role,
             ecr_repository=ecr_repository,
             container_runtime=container_runtime,
             auto_create_ecr=auto_create_ecr,

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/src/bedrock_agentcore_starter_toolkit/notebook/runtime/bedrock_agentcore.py

@@ -35,6 +35,7 @@ class Runtime:
         self,
         entrypoint: str,
         execution_role: Optional[str] = None,
+        code_build_execution_role: Optional[str] = None,
         agent_name: Optional[str] = None,
         requirements: Optional[List[str]] = None,
         requirements_file: Optional[str] = None,
@@ -53,6 +54,7 @@ class Runtime:
             entrypoint: Path to Python file with optional Bedrock AgentCore name
                 (e.g., "handler.py" or "handler.py:bedrock_agentcore")
             execution_role: AWS IAM execution role ARN or name (optional if auto_create_execution_role=True)
+            code_build_execution_role: Optional separate CodeBuild execution role ARN or name
             agent_name: name of the agent
             requirements: Optional list of requirements to generate requirements.txt
             requirements_file: Optional path to existing requirements file
@@ -109,6 +111,7 @@ class Runtime:
             entrypoint_path=Path(file_path),
             auto_create_execution_role=auto_create_execution_role,
             execution_role=execution_role,
+            code_build_execution_role=code_build_execution_role,
             ecr_repository=ecr_repository,
             container_runtime=container_runtime,
             auto_create_ecr=auto_create_ecr,

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/src/bedrock_agentcore_starter_toolkit/operations/runtime/configure.py

@@ -12,6 +12,7 @@ from ...utils.runtime.schema import (
     AWSConfig,
     BedrockAgentCoreAgentSchema,
     BedrockAgentCoreDeploymentInfo,
+    CodeBuildConfig,
     NetworkConfiguration,
     ObservabilityConfig,
     ProtocolConfiguration,
@@ -25,6 +26,7 @@ def configure_bedrock_agentcore(
     agent_name: str,
     entrypoint_path: Path,
     execution_role: Optional[str] = None,
+    code_build_execution_role: Optional[str] = None,
     ecr_repository: Optional[str] = None,
     container_runtime: Optional[str] = None,
     auto_create_ecr: bool = True,
@@ -43,6 +45,7 @@ def configure_bedrock_agentcore(
         agent_name: name of the agent,
         entrypoint_path: Path to the entrypoint file
         execution_role: AWS execution role ARN or name (auto-created if not provided)
+        code_build_execution_role: CodeBuild execution role ARN or name (uses execution_role if not provided)
         ecr_repository: ECR repository URI
         container_runtime: Container runtime to use
         auto_create_ecr: Whether to auto-create ECR repository
@@ -109,6 +112,24 @@ def configure_bedrock_agentcore(
             else:
                 log.debug("No execution role provided and auto-create disabled")
 
+    # Handle CodeBuild execution role - use separate role if provided, otherwise use execution_role
+    codebuild_execution_role_arn = None
+    if code_build_execution_role:
+        # User provided a separate CodeBuild role
+        if code_build_execution_role.startswith("arn:aws:iam::"):
+            codebuild_execution_role_arn = code_build_execution_role
+        else:
+            codebuild_execution_role_arn = f"arn:aws:iam::{account_id}:role/{code_build_execution_role}"
+
+        if verbose:
+            log.debug("Using separate CodeBuild execution role: %s", codebuild_execution_role_arn)
+    else:
+        # No separate CodeBuild role provided - use None
+        codebuild_execution_role_arn = None
+
+        if verbose and execution_role_arn:
+            log.debug("Using same role for CodeBuild: %s", codebuild_execution_role_arn)
+
     # Generate Dockerfile and .dockerignore
     bedrock_agentcore_name = None
     # Try to find the variable name for the Bedrock AgentCore instance in the file
@@ -195,6 +216,9 @@ def configure_bedrock_agentcore(
             observability=ObservabilityConfig(enabled=enable_observability),
         ),
         bedrock_agentcore=BedrockAgentCoreDeploymentInfo(),
+        codebuild=CodeBuildConfig(
+            execution_role=codebuild_execution_role_arn,
+        ),
         authorizer_configuration=authorizer_configuration,
         request_header_configuration=request_header_configuration,
     )

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/src/bedrock_agentcore_starter_toolkit/utils/runtime/templates/Dockerfile.j2

@@ -8,7 +8,7 @@ ENV UV_SYSTEM_PYTHON=1 UV_COMPILE_BYTECODE=1
 {% if dependencies_install_path %}
 COPY {{ dependencies_install_path }} {{ dependencies_install_path }}
 # Install from pyproject.toml directory
-RUN uv pip install {{ dependencies_install_path }}
+RUN cd {{ dependencies_install_path }} && uv pip install .
 {% else %}
 COPY {{ dependencies_file }} {{ dependencies_file }}
 # Install from requirements file

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/tests/cli/runtime/test_commands.py

@@ -126,6 +126,51 @@ def handler(payload):
             call_args = mock_configure.call_args
             assert call_args[1]["authorizer_configuration"] == oauth_config
 
+    def test_configure_with_code_build_execution_role(self, tmp_path):
+        """Test configure command with CodeBuild execution role."""
+        agent_file = tmp_path / "test_agent.py"
+        agent_file.write_text("from bedrock_agentcore.runtime import BedrockAgentCoreApp\napp = BedrockAgentCoreApp()")
+
+        with (
+            patch(
+                "bedrock_agentcore_starter_toolkit.cli.runtime.commands.configure_bedrock_agentcore"
+            ) as mock_configure,
+            patch("bedrock_agentcore_starter_toolkit.cli.runtime.commands.parse_entrypoint") as mock_parse,
+            patch(
+                "bedrock_agentcore_starter_toolkit.cli.runtime.commands._handle_requirements_file_display"
+            ) as mock_req_display,
+            patch("bedrock_agentcore_starter_toolkit.cli.common.prompt") as mock_prompt,
+        ):
+            mock_parse.return_value = (str(agent_file), "bedrock_agentcore")
+            mock_req_display.return_value = tmp_path / "requirements.txt"
+            mock_prompt.return_value = "no"
+
+            mock_result = Mock()
+            mock_result.runtime = "docker"
+            mock_result.region = "us-west-2"
+            mock_result.account_id = "123456789012"
+            mock_result.execution_role = "arn:aws:iam::123456789012:role/ExecutionRole"
+            mock_result.config_path = tmp_path / ".bedrock_agentcore.yaml"
+            mock_configure.return_value = mock_result
+
+            result = self.runner.invoke(
+                app,
+                [
+                    "configure",
+                    "--entrypoint",
+                    str(agent_file),
+                    "--execution-role",
+                    "ExecutionRole",
+                    "--code-build-execution-role",
+                    "CodeBuildRole",
+                ],
+            )
+
+            assert result.exit_code == 0
+            # Verify CodeBuild execution role was passed
+            call_args = mock_configure.call_args
+            assert call_args[1]["code_build_execution_role"] == "CodeBuildRole"
+
     def test_configure_with_invalid_protocol(self, tmp_path):
         agent_file = tmp_path / "test_agent.py"
 

{bedrock_agentcore_starter_toolkit-0.1.13 → bedrock_agentcore_starter_toolkit-0.1.14}/tests/notebook/runtime/test_bedrock_agentcore.py

@@ -83,6 +83,33 @@ def handler(payload):
             assert "boto3" in content
             assert "pandas" in content
 
+    def test_configure_with_code_build_execution_role(self, tmp_path):
+        """Test configuration with CodeBuild execution role."""
+        agent_file = tmp_path / "test_agent.py"
+        agent_file.write_text("from bedrock_agentcore.runtime import BedrockAgentCoreApp\napp = BedrockAgentCoreApp()")
+
+        bedrock_agentcore = Runtime()
+
+        with (
+            patch(
+                "bedrock_agentcore_starter_toolkit.notebook.runtime.bedrock_agentcore.configure_bedrock_agentcore"
+            ) as mock_configure,
+        ):
+            mock_result = Mock()
+            mock_result.config_path = tmp_path / ".bedrock_agentcore.yaml"
+            mock_configure.return_value = mock_result
+
+            bedrock_agentcore.configure(
+                entrypoint=str(agent_file),
+                execution_role="ExecutionRole",
+                code_build_execution_role="CodeBuildRole",
+            )
+
+            # Verify configure was called with CodeBuild execution role
+            mock_configure.assert_called_once()
+            args, kwargs = mock_configure.call_args
+            assert kwargs["code_build_execution_role"] == "CodeBuildRole"
+
     def test_launch_without_config(self):
         """Test launch fails when not configured."""
         bedrock_agentcore = Runtime()